JP3041895B2 - Noise source for electronic musical instruments - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a noise sound source device for generating a more realistic and expressive musical tone in an electronic musical instrument.

[Conventional technology]

In an electronic musical instrument, a desired tone is generated by generating a sound source signal having a pitch corresponding to a pressed key and performing necessary processing such as filtering and envelope control on the sound source signal.

By the way, the performance sounds of musical instruments include not only pure musical sounds but also so-called non-musical noise components such as sibilance, repellent sounds, whistling sounds, and unanalyzable sounds. It has a great effect on the texture of the sound. Therefore, in order to generate a realistic and highly expressive musical tone closer to a raw sound with an electronic musical instrument, it is necessary to add this noise component to the musical tone.

Conventionally, as a noise sound source device for generating such a noise component, a noise waveform extracted from a performance sound is stored in a memory, and the noise waveform is read out according to a musical sound generation operation, or There is known a device that uses white noise composed of an analog signal and adjusts and adds the signal level.

[Problems to be solved by the invention]

However, in the case of a noise sound source device in which a noise waveform is directly stored, the same preset noise is always added, and in the case of a noise sound source device using white noise, the tim However, there is a problem that the tone generated in any case is monotonous and lacks a natural feeling because white noise having no characteristic is added to the tone.

The present invention has been made based on the above circumstances, and an object of the present invention is to provide a noise sound source device capable of generating a more realistic and expressive musical tone.

[Means for solving the problem]

In order to achieve the above object, a first noise sound source device includes:
Noise signal generating means for generating a random noise signal; and a noise component of the original sound obtained based on an analysis result of each sample by sampling a sustain portion of the original sound to be synthesized a plurality of times under predetermined conditions. Parameter storage means for storing a time-varying parameter signal for reproducing the shape of each tone for each tone, and reading out the parameter signal of the tone according to the tone generation instruction in response to the tone generation instruction;
Noise shape control means for controlling the shape of the noise signal generated by the noise signal generation means based on the parameter signal read from the parameter storage means.

The second noise source device includes a noise signal generating means for generating a random noise signal, and a plurality of samples of a sustain portion of an original sound to be synthesized under predetermined conditions. A time-varying parameter signal that reproduces the peak characteristic of the spectrum of the noise component of the original sound, which is obtained based on the tone signal, is stored for each tone color, and the parameter of the tone color according to the tone generation instruction is stored in response to the tone generation instruction. Parameter storage means from which a signal is read;
Noise peak control means for controlling a peak characteristic of a spectrum of a noise signal generated from the noise signal generation means based on the parameter signal read from the parameter storage means.

(Operation)

In the case of the first noise sound source device, when an instruction to generate a musical tone is given, parameter signals for controlling the shape of the noise signal are sequentially read from the parameter storage means in accordance with the tone color according to the instruction to generate the musical tone. The noise shape control means changes the shape of the noise signal according to the time-varying parameter signal output from the parameter storage means. The noise shape control means can be constituted by, for example, a digital filter. By controlling the peak frequency characteristic and the bandwidth characteristic of the digital filter based on the parameter signal, the frequency component of the noise signal input to the digital filter, that is, The shape of the noise signal can be controlled over time. In this case, the parameter signal for controlling the shape of the noise signal is stored in advance in the parameter storage means. This parameter signal is obtained by sampling the sustain part of the original sound to be synthesized a plurality of times under predetermined conditions. It is a temporally varying form that reproduces the shape of the noise component of the original sound obtained based on the analysis result of each sample, and is stored for each tone. For example, it corresponds to a complicated characteristic of a noise component included in a target natural musical instrument sound or the like. As a result, a noise signal controlled to a desired state while having randomness can be obtained. In addition, the tone shape unique to the tone can be reproduced for each tone, and the noise shape is parameterized by using the sustain part of the original sound. Can be reproduced. Therefore, by adding this noise signal to a normal tone signal,
Very realistic and natural musical sounds can be generated.

In the case of the second noise sound source device, when an instruction to generate a musical tone is given, parameter signals for controlling the peak characteristic of the spectrum of the noise component are successively stored from the parameter storage means in correspondence with the timbre according to the instruction to generate the musical tone. Is read. The noise peak control means controls a peak characteristic of a spectrum included in the noise signal generated by the noise signal generation means according to the parameter signal from the parameter generation means. This noise peak control means can be constituted by, for example, a digital filter. By controlling a peak frequency characteristic and a bandwidth characteristic of the digital filter based on a parameter signal, the noise peak control means is included in the noise signal input to the digital filter. The peak characteristic of the frequency component can be controlled. This parameter signal is obtained by sampling the sustain part of the original sound to be synthesized a plurality of times under predetermined conditions and reproducing the peak characteristic of the spectrum of the noise component of the original sound obtained based on the analysis result of each sample. And stored for each tone color. As a result, a noise signal having a spectrum that is formant-controlled to a desired state similar to the noise component included in a natural musical instrument or the like while having randomness can be obtained. Also, since the spectral peak of noise specific to the tone can be reproduced for each tone, and the spectral peak of the noise is parameterized using the sustain part of the original sound, the noise unique to the tone is not merely mechanical noise. The components can be reliably reproduced. Therefore, by adding this noise signal to a normal tone signal, a very realistic and natural tone can be generated.

〔Example〕

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 shows a first embodiment of the noise source device of the present invention. A random noise signal composed of a digital signal generated by the random noise generator 1 is sent to the bandpass filter 2 as an original sound for noise generation. The band-pass filter 2 is a digital filter capable of variably controlling the filter characteristics, that is, the center frequency f ₀ of the pass band and the pass band width B. A noise signal passing through this filter is input to the multiplier 3. The multiplier 3 controls the envelope of the noise signal by multiplying the noise signal transmitted from the band-pass filter 2 by the noise envelope characteristic data transmitted from the noise envelope characteristic memory 6.

Noise frequency characteristic memory 4 comprises a semiconductor memory such as ROM or RAM, sampling the playing sound of the musical instrument for each strength of the key touch for each key, the time course curves of the peak frequency f ₀ of the noise component of each key It is stored as a noise frequency characteristic table FF (t) for each key touch. FIG. 2 shows an example of the noise frequency characteristic table FF (t). FIG. 2 shows a noise frequency characteristic table for each key touch when the intensity of the initial key touch for one key is changed to n stages, and the noise frequency characteristic memory 4 stores such a characteristic table. Is stored for each key.

The noise bandwidth characteristic memory 5 is composed of a semiconductor memory such as a ROM or a RAM, and samples the performance sound of a musical instrument for each key touch strength for each key, and changes the noise component bandwidth over time, for example, the peak frequency. From f ₀ its amplitude is 3 dB Is obtained for each key touch of each key, and the bandwidth B is converted into a filter selectivity Q and stored as a noise bandwidth characteristic table FQ (t). I have. This noise bandwidth characteristic table FQ (t)
An example is shown in FIG. FIG. 3 shows a noise bandwidth characteristic table for each key touch when the intensity of the initial key touch for one key is changed to n stages. A characteristic table is stored for each key.

Note that between the peak frequency f ₀ of the noise component and the bandwidth B and the selectivity Q of the filter, Q = f ₀ / B
And B can be easily converted to Q.

The noise envelope characteristic memory 6 is composed of a semiconductor memory such as a ROM or a RAM, and samples the performance sound of the musical instrument for each key touch strength of each key, and calculates a time-dependent change curve of the amplitude of the noise component for each key. It is stored as a noise envelope characteristic table NE (t) for each touch. This noise envelope characteristic table NE (t)
FIG. 4 shows an example. FIG. 4 shows a noise envelope characteristic table for each key touch when the intensity of the initial key touch for one key is changed in n stages. The noise envelope characteristic memory 6 stores such a characteristic table. Is stored for each key.

The read-out circuits 7, 8, and 9 receive the tone generation instruction information, that is, a key-on signal, a key code indicating a pressed key (ie, a pitch of a tone to be generated) and key touch data as input information for an address. Select a key (key code) pressed at that time from 4, 5, and 6 and a characteristic table corresponding to the key touch intensity (key touch data), and specify the data of each characteristic table in response to a key-on signal. Is a circuit for sequentially reading out in synchronization with the clock. Noise frequency characteristic table FF read from noise frequency characteristic memory 4 and noise bandwidth characteristic memory 5
(T) and the data of the noise bandwidth characteristic table FQ (t) are sent to the band-pass filter 2, and the data of the noise envelope characteristic table NE (t) read from the noise envelope characteristic memory 6 are multiplied. Sent to the vessel 3.

The band-pass filter 2 has a filter characteristic according to the data of the noise frequency characteristic table FF (t) and the noise bandwidth characteristic table FQ (t) sent from the noise frequency characteristic memory 4 and the noise bandwidth characteristic memory 5.
That is, the center frequency f ₀ of the pass band and the pass band width B
(Filter selectivity Q) is variably controlled. Therefore,
In the random noise signal input to the band-pass filter 2, the peak frequency f ₀ of the noise component follows a change curve as illustrated in FIG. 2, and the bandwidth B of the noise component is illustrated in FIG. According to such a change curve, the noise signal is processed into a noise signal which changes with time, and is sent to the multiplier 3.

On the other hand, the data of the noise envelope characteristic table NE (t) read from the noise envelope characteristic memory 6 is sent to the multiplier 3 as described above. The multiplier 3
The data of the noise envelope characteristic table NE (t) is multiplied by the noise signal sent from the band-pass filter 2, and the envelope is variably controlled. Therefore, the noise signal input to the multiplier 3 is a noise signal whose envelope changes with time according to a change curve as exemplified in FIG.

As a result, the noise signal output from the multiplier 3 is the difference between the key pressed at that time by the peak frequency f ₀ of the noise component in the noise signal, its bandwidth B, and its envelope, and the key touch at that time. , The noise signal changes adaptively with time, and becomes a noise signal very similar to the noise contained in the actual performance sound of the musical instrument.

FIG. 5 shows a configuration example of a measurement circuit of the noise frequency characteristic table FF (t), the noise bandwidth characteristic FQ (t), and the noise envelope characteristic table NE (t) stored in the memories 4, 5, and 6. In the example of FIG. 5, a piano 10 is used as a live musical instrument for sampling, and each key of the piano 10 has a solenoid 11 for tapping and a driver for driving the solenoid.
A key 13 can be freely touched while freely changing the key touch strength by controlling the driver 12 with the microcomputer 13. Then, the pronunciation of the piano 10 at the time of this keying is performed by the microphone 14 and the amplifier 15.
The digital data is converted by the A / D converter 16 and stored in the internal RAM of the microcomputer 13, and the stored raw sound sampling data is subjected to the fast Fourier transform (FF
By using T) or the like, a noise frequency characteristic table FF (t), a noise bandwidth characteristic table FQ (t), and a noise envelope characteristic table NE (t) are obtained for each key touch of each key. It was done.

That is, the operation of the measurement circuit shown in FIG. 5 will be described with reference to the flowchart shown in FIG. 6. First, the microcomputer 13 specifies the key of the lowest tone as the key code KCD.
After setting CD = 0 (step S1) and setting the key touch intensity TD to TD = 0 designating the weakest sound (step S2), the driver 12 is controlled to control the key of the key code KCD = 0, that is, the lowest key. Touch the key of the sound with the weakest key touch.
The raw sound of the piano 10 resulting from this keying is picked up by the microphone 14 and the amplifier 15, converted into a digital signal by the A / D converter 16, and stored in the internal RAM of the microcomputer 13. And
Using the stored sampling data, the detailed operation is performed by a noise parameter measurement routine of step S3 described later, the noise frequency characteristic table FF (t) and the noise bandwidth characteristic table FQ for the key touch intensity at that time.
(T) and noise envelope characteristic table NE (t)
Ask for.

The above process is repeated for one key while changing the key touch intensity TD thereof in eight steps (TD = 0 to 7) (steps S4 and S5). Further, this is repeated for all keys from the lowest key to the highest key. Execute sequentially for keys (Steps S6, S
7).

As a result, the noise frequency characteristic table FF (t) and the noise bandwidth characteristic table FQ for eight key touch intensities are provided for each key from the lowest key to the highest key.
(T) and noise envelope characteristic table NE (t)
Thus, the microcomputer 13 writes the obtained characteristic table into the noise frequency characteristic memory 4, the noise bandwidth characteristic memory 5, and the noise envelope characteristic memory 6, respectively.

FIG. 7 is a flowchart showing details of the noise parameter measurement routine (step S3) in FIG. In FIG. 6, each time a key is touched N times, each characteristic table is obtained as an average of the N times of keying. In steps S11, S18, and S19, the number i of keying is monitored. doing.

First, drive the solenoid 11 and key code on the piano 10
A sound corresponding to the KCD and the key touch intensity TD is generated, and the sound is sampled (step S12). Then, a noise component is extracted by using the waveform of the sustain portion of the sampled raw sound (step S13). Then, the envelope e (t) of the noise component is extracted and normalized (step S14). By adding the obtained envelope e (t) to NE (t) up to the previous time, a noise envelope characteristic table NE (t) by averaging N times is obtained (step S15).

On the other hand, the normalized noise component is subjected to Fast Fourier Transform (FFT) for each fixed interval Δt, and its frequency characteristic f (t,
v) (step S16), and the obtained frequency characteristic f
By adding (t, ν) to the frequency characteristic F (t, ν) obtained up to the previous time, the frequency characteristic F by N keystrokes is obtained.
(T, v) is obtained (step S17).

When N keystrokes have been completed for one key touch intensity (step S18), the frequency characteristics F (t, ν) for the N times obtained in step S17 are analyzed to determine the noise frequency characteristics at the key touch intensity at that time. Table FF
(T) and a noise bandwidth characteristic table FQ (t) are obtained (step S20). Then, the obtained noise frequency characteristic table FF (t) is written in the noise frequency characteristic memory 4, the noise bandwidth characteristic table FQ (t) is written in the noise bandwidth characteristic memory 5, and the noise obtained in step S15 is written. The envelope characteristic table NE (t) is written in the noise envelope characteristic memory 6 (step S21). Further, if necessary, a repetition loop of the noise component for the sustain portion of the noise signal is set (step S22).

FIG. 8 shows an example of a musical sound generation device configured using the noise source device of the first embodiment (FIG. 1). In the figure, the block denoted by reference numeral 23 is the noise source device of FIG. In this example, a tone signal is generated by the well-known PCM tone generator 22. The normal noise source device 23
The PCM tone generator 22 has a multi-channel configuration, and is capable of simultaneously generating a tone signal and its noise signal in a time-division manner for a plurality of keys. The noise sound source device 23 is basically the same as that shown in FIG. 1, but in this case, each of the memories 4, 5, and 6 is provided for each tone and each key (each pitch). ) And a table FF (t), FQ (t), NE (t) for each key touch.

When the timbre switch 20 is operated and a predetermined musical instrument, for example, a piano is selected, a timbre designation signal designating a piano tone is sent from the timbre switch detection circuit 21 to the PCM tone generator 22 and the noise tone generator 23, and the respective tone generators The circuits required to generate the piano sound are selected.

When the keyboard 17 is depressed, necessary keys such as a key-on signal, a key code indicating a pitch, and a key touch code indicating a key touch intensity in FIG. Key information is sent. The channel assignment circuit 19 determines which key sound or the like is to be emitted on the channel according to the key press information, and converts the channel designation code, key-on signal, key code, and key touch code to the PC.
It is sent to the M sound source device 22 and the noise sound source device 23.

The PCM tone generator 22 corresponds to the key pitch and timbre of the key code at that time based on the data sent from the channel assignment circuit 19 and the timbre switch detection circuit 21,
Further, a tone signal having a predetermined envelope as shown in FIG. 9A corresponding to the key-on length and the key touch intensity at that time is generated and output.

On the other hand, as described with reference to FIG. 1, the noise sound source device 23 corresponds to the key and the tone of the key code at that time based on the data transmitted from the channel assignment circuit 19 and the tone color switch detection circuit 21. A predetermined noise signal having a noise frequency characteristic, a noise bandwidth characteristic, and a noise envelope characteristic corresponding to the key touch intensity at that time is generated and output as shown in FIG. 9B.

After the tone signal and the noise signal of each channel generated in a time-division manner by the PCM tone generator 22 and the noise tone generator 23 are added in an adder 24, the tone signals of each channel are superimposed in a channel accumulator 25. , Via the D / A converter 26 to the sound system.

When the sustain portion S of the musical tone is long as shown in FIG. 9 (a), the sustain portion of the noise signal repeats the same noise pattern in a loop as shown in FIG. 9 (b). What should I do?

FIG. 10 shows a second embodiment of the noise source device of the present invention. In the second embodiment, a noise signal generating circuit
The noise signals output from the respective noise signal generation circuits are added in a system parallel manner and added by an adder 29 and output as a final noise signal. With such a configuration, as shown in FIG.
Noise signals having different peak frequencies f _{0a to} f _0d can be independently generated for each system, and by combining these noise signals, a noise signal closer to actual noise can be generated. In FIG. 10, 2a to 2d are band-pass filters, and 3a to 3d are multipliers. Also, the parameter generator 28
Each of a to 28d has a built-in characteristic memory 4 to 6 in FIG. 1 and a readout circuit 7 to 9 in FIG. 1, and a unique noise frequency characteristic table FF (t ) 1 to FF (t) 4, a band-pass filter 2 a corresponding to the data of the noise bandwidth characteristic tables FQ (t) 1 to FQ (t) 4 and the noise envelope characteristic tables NE (t) 1 to NE (t) 4. ~ 2d and multipliers 3a ~ 3d
To be sent to.

FIG. 12 shows a third embodiment of the noise source device of the present invention. The third embodiment is configured as a noise sound source device particularly suitable for a tone generator using an FM sound source. In the figure, the portion indicated by reference numeral 30 constitutes the above-described noise source device of the present invention, and comprises a random noise generator 1, a low-pass filter 2e, a multiplier 3, characteristic memories 4 to 6, and a read circuit 7a. I have. Then, in addition to the configuration of the noise source device 30 of the present invention, an envelope generator
31 and a multiplier 32 are added to provide a noise sound source device suitable for an FM sound source. The noise signal output from the multiplier 3 is transmitted to the FM sound source apparatus shown in FIG.
The noise signal branched from the low-pass filter 2e is sent to the adder 24 in FIG. 13 after the envelope is modulated by the envelope signal output from the envelope generator 31 in the multiplier 32. .

FIG. 13 shows an example of a musical sound generation device configured using the noise source device of FIG. The tone generator shown in FIG. 13 employs the FM tone generator 34 as a tone signal tone generator as described above. The same circuits as those in FIG. 8 are denoted by the same reference numerals, and description thereof will be omitted.

As is well known, the FM sound source device 34 is configured to use a module called an operator and to modulate phase data for generating a sine wave (address signal of a sine table) with a modulation signal in the operator. By changing the number of connections of the operator, the connection method, and the feedback method, a desired tone signal is generated. The noise signal output from the multiplier 3 in FIG. 12 is input to the FM sound source device 34 as one of modulation data for FM modulation. Also, the twelfth
The noise signal output from the multiplier 32 in the figure is sent to the adder 24, and is added as a noise signal to the tone signal generated by the FM tone generator 34.

In general, the frequency spectrum of an FM sound source is mostly composed of a fundamental tone and its harmonic components, as shown in FIG. 14 (a), and has very few non-harmonic components. In the musical tone signal having the spectrum as shown in FIG. 14 (a), the sound is too monotonous. Therefore, as shown in FIG. 14 (b), the tone signal needs to be generated as a tone signal having a swelling of non-overtone components around the fundamental and overtone components.
Therefore, in order to generate a tone signal having a spectrum as shown in FIG. 14 (b), each of the characteristic memories 4 to 4 in FIG.
It is desirable to store each characteristic table which gives the hatched frequency components in FIG. 14 (b).

FIG. 15 shows a characteristic table FF suitable for adding a frequency component as shown by hatching in FIG. 14 (b).
An example of a noise parameter measurement method for obtaining (t), FQ (t), and NE (t) will be described. The flowchart of FIG. 15 uses the measuring circuit of FIG. 10 as a measuring circuit, and taps each key only once for each key touch, and obtains a noise characteristic table for each key touch from sampling data obtained at this time. Is what you get.

That is, the flowchart of FIG. 15 will be briefly described. First, the solenoid 11 is driven, and the key code KC
A sound corresponding to D and the key touch intensity TD is generated, and the sound is sampled (step S30). Then, after the envelope is removed from the sampled waveform to obtain a normalized waveform having a constant amplitude (step S31), a fast Fourier transform (FFT) is performed on the waveform for each fixed interval t, and its frequency characteristic is calculated. (Steps S32 and S33).

From the frequency characteristics obtained as described above, as shown in FIG. 16 (a), the frequency component of the fundamental wave portion is extracted in a form in which it is preserved up to the foot portion thereof (step S34),
The frequency axis is transformed so that the center position of this frequency component is at the 0 Hz position, and the frequency components on the left and right sides of the center position are added to obtain a frequency characteristic of only one side suitable for a low-pass filter as shown in FIG. 16 (b). (Step S35).
Then, the cutoff frequency, selectivity Q, and attenuation level value of the low-pass filter 2e that simulates the frequency characteristic of FIG. 16B are calculated, and FF (t), FQ (t ) And NE (t) are obtained (step S36).

The noise frequency characteristic table FF (t) that gives a time-dependent change curve of the frequency component of the fundamental wave portion at the key touch strength at that time by sequentially performing the above processing for each section t up to the last section (steps S37 and S38). , A noise bandwidth characteristic table FQ (t) and a noise envelope characteristic table NE (t) are obtained. By performing such processing for each key touch intensity for each key of various musical instruments, the characteristic tables FF (t), FQ (t), and NE (t) for each key touch intensity for each key for each musical instrument ) Are obtained respectively. Therefore, the obtained characteristic tables may be written in the characteristic memories 4 to 6 in FIG.

If the noise signal generated by the noise source device 33 incorporating the respective characteristic tables obtained as described above is given to the FM source device 34 as one of the modulation data for FM modulation, the FM
The tone generator 34 generates a natural tone signal including a fundamental wave and a non-overtone component having a certain width around the overtone, as shown in FIG. 14 (b).

In each of the above embodiments, each of the characteristic memories 4-6
For each key (each pitch), the table FF (t), FQ
Although (t) and NE (t) are stored, these tables may be stored for each key range (each range) instead of each key.

〔The invention's effect〕

As is apparent from the above description, in the case of the noise source device according to the first aspect, the sustain part of the original sound to be synthesized, which is stored in the parameter storage means, is sampled a plurality of times under predetermined conditions, and each time. Since the shape of the random noise signal is controlled according to the parameter signal for each time-varying tone that reproduces the shape of the noise component of the original sound obtained based on the analysis result of the sample, the shape of the noise signal Can be controlled over time and in a complicated manner, and a noise signal very similar to a noise component included in a natural musical instrument or the like can be obtained. In addition, the tone shape unique to the tone can be reproduced for each tone, and the noise shape is parameterized using the sustain part of the original sound. Can be reproduced. Therefore, if the noise signal thus obtained is added to the musical tone signal, it is possible to generate a musical tone that is more realistic and expressive than before.

In the case of the noise source device according to claim 2,
The peak characteristic of the spectrum of the random noise signal, the sustain part of the original sound to be synthesized is sampled a plurality of times under predetermined conditions, and the spectrum of the noise component of the original sound obtained based on the analysis result of each sample The characteristic of the random noise signal can be easily controlled to a desired state by controlling according to the parameter signal for each time-varying timbre that reproduces the peak characteristic of the noise. A noise signal having a spectrum very similar to the component can be easily obtained. Also, since the spectral peak of noise specific to the tone can be reproduced for each tone, and the spectral peak of the noise is parameterized using the sustain part of the original sound, the noise unique to the tone is not merely mechanical noise. The components can be reliably reproduced. Therefore, if the noise signal thus obtained is added to the musical tone signal, it is possible to generate a musical tone that is more realistic and expressive than before.

[Brief description of the drawings]

FIG. 1 is a block diagram of a first embodiment of a noise source device of the present invention, FIG. 2 is a diagram showing an example of a noise frequency characteristic table for the present invention, and FIG. 3 is a noise band for the present invention. FIG. 4 is a diagram showing an example of a width characteristic table, FIG. 4 is a diagram showing an example of a noise envelope characteristic table for the present invention, FIG. 5 is a configuration diagram of a noise parameter measuring circuit in the first embodiment, FIG. 7 is a flowchart of the processing operation of the measurement circuit, FIG. 7 is a detailed flowchart of a noise parameter measurement routine in FIG. 6, and FIG. 8 is a flowchart of a tone generation device configured using the noise source device of the first embodiment. FIG. 9 is a block diagram showing an example, FIG. 9 is a diagram showing a waveform example of a tone signal and a noise signal generated by the tone generator of FIG. 8, and FIG. 10 is a diagram of a second embodiment of the noise source device of the present invention. Block diagram, Fig. 11 FIG. 12 is an explanatory diagram of a noise signal generated in the second embodiment, FIG. 12 is a block diagram of a third embodiment of the noise source device of the present invention, and FIG. 13 uses the noise source device of the third embodiment. FIG. 14 is a block diagram showing an example of a musical sound generating device configured with the above. FIG. 14 is a diagram showing frequency characteristics of a musical sound signal generated by an FM sound source. FIG. FIG. 16 is a flowchart of a measurement method, and FIG. 15 is a diagram showing frequency characteristics of a fundamental wave component. 1 random noise generator, 2, 2a to 2d bandpass filter, 2e low pass filter, 3, 3a to 3e multiplier, 4 noise frequency characteristic memory, 5 noise bandwidth characteristic Memory 6, Noise envelope characteristic memory,
FF (t): Noise frequency characteristic table, FQ (t):
Noise bandwidth characteristic table, NE (t) ... Noise envelope characteristic table.

Continuation of the front page (72) Inventor Masahiro Shimizu 10-1 Nakazawa-cho, Hamamatsu-shi, Shizuoka Prefecture Inside Yamaha Corporation (56) References JP-A-1-209497 (JP, A) JP-A-55-48791 (JP) JP-A-54-145121 (JP, A) JP-A-61-259300 (JP, A) JP-A-62-194289 (JP, A) JP-A-56-146197 (JP, A) 56-144492 (JP, A) Japanese Utility Model Showa 57-100797 (JP, U) JP-B 61-26080 (JP, B2) JP-B 3-50280 (JP, B2) JP-B 62-9919 (JP, U.S.A.) B2) JP 4-29080 (JP, B2) JP 3-40399 (JP, B2) JP 2-43199 (JP, B2) (58) Fields surveyed (Int. Cl. ⁷ , DB name) ) G10H 1/00-7/12

Claims (2)

(57) [Claims] 1. A noise signal generating means for generating a random noise signal, and a sustain portion of an original sound to be synthesized is sampled a plurality of times under predetermined conditions, and is obtained based on an analysis result of each sample. Parameter storage means for storing a time-varying parameter signal for reproducing the shape of the noise component of the original sound for each timbre, and reading out the parameter signal of the timbre according to the tone generation instruction in response to the tone generation instruction; and A noise shape control means for controlling a shape of a noise signal generated from the noise signal generation means based on a parameter signal read from the parameter storage means. . 2. A noise signal generating means for generating a random noise signal, and a sustain part of an original sound to be synthesized is sampled a plurality of times under predetermined conditions, and is obtained based on an analysis result of each sample. A parameter for storing a time-varying parameter signal for reproducing the peak characteristic of the spectrum of the noise component of the original sound for each timbre, and reading the parameter signal of the timbre according to the tone generation instruction in response to the tone generation instruction Storage means; and a noise peak control means for controlling a peak characteristic of a spectrum of a noise signal generated from the noise signal generation means based on the parameter signal read from the parameter storage means. Noise source device for electronic musical instruments.